solrium 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/LICENSE.txt +712 -0
- data/README.markdown +51 -0
- data/Rakefile +24 -0
- data/VERSION.yml +4 -0
- data/bin/solr +9 -0
- data/lib/jars/jetty/ant-1.6.5.jar +0 -0
- data/lib/jars/jetty/commons-codec-1.3.jar +0 -0
- data/lib/jars/jetty/commons-fileupload-1.2.jar +0 -0
- data/lib/jars/jetty/jetty-7.0.0.pre5.jar +0 -0
- data/lib/jars/jetty/jetty-util-7.0.0.pre5.jar +0 -0
- data/lib/jars/jetty/jsp-2.1.jar +0 -0
- data/lib/jars/jetty/jsp-api-2.1.jar +0 -0
- data/lib/jars/jetty/servlet-api-3.0.pre4.jar +0 -0
- data/lib/jars/lucene/lucene-analyzers-2.4-dev.jar +0 -0
- data/lib/jars/lucene/lucene-core-2.4-dev.jar +0 -0
- data/lib/jars/lucene/lucene-highlighter-2.4-dev.jar +0 -0
- data/lib/jars/lucene/lucene-memory-2.4-dev.jar +0 -0
- data/lib/jars/lucene/lucene-queries-2.4-dev.jar +0 -0
- data/lib/jars/lucene/lucene-snowball-2.4-dev.jar +0 -0
- data/lib/jars/lucene/lucene-spellchecker-2.4-dev.jar +0 -0
- data/lib/jars/solr/apache-solr-common-1.3.0.jar +0 -0
- data/lib/jars/solr/apache-solr-core-1.3.0.jar +0 -0
- data/lib/jars/solr/commons-io-1.2.jar +0 -0
- data/lib/jars/solr/stax-1.2.0-dev.jar +0 -0
- data/lib/jars/solr/stax-api-1.0.jar +0 -0
- data/lib/jars/solr/stax-utils.jar +0 -0
- data/lib/jars/solr/xpp3-1.1.3.4.O.jar +0 -0
- data/lib/jetty.rb +83 -0
- data/lib/lucene.rb +33 -0
- data/lib/queries.rb +60 -0
- data/lib/remote.rb +22 -0
- data/lib/solr.rb +69 -0
- data/lib/solr_ext.rb +81 -0
- data/lib/webapp/admin/_info.jsp +110 -0
- data/lib/webapp/admin/action.jsp +116 -0
- data/lib/webapp/admin/analysis.jsp +456 -0
- data/lib/webapp/admin/analysis.xsl +179 -0
- data/lib/webapp/admin/distributiondump.jsp +158 -0
- data/lib/webapp/admin/favicon.ico +0 -0
- data/lib/webapp/admin/form.jsp +129 -0
- data/lib/webapp/admin/get-file.jsp +72 -0
- data/lib/webapp/admin/get-properties.jsp +24 -0
- data/lib/webapp/admin/header.jsp +41 -0
- data/lib/webapp/admin/index.jsp +154 -0
- data/lib/webapp/admin/jquery-1.2.3.min.js +32 -0
- data/lib/webapp/admin/logging.jsp +50 -0
- data/lib/webapp/admin/logging.xsl +91 -0
- data/lib/webapp/admin/meta.xsl +34 -0
- data/lib/webapp/admin/ping.jsp +65 -0
- data/lib/webapp/admin/ping.xsl +71 -0
- data/lib/webapp/admin/raw-schema.jsp +38 -0
- data/lib/webapp/admin/registry.jsp +107 -0
- data/lib/webapp/admin/registry.xsl +287 -0
- data/lib/webapp/admin/schema.jsp +661 -0
- data/lib/webapp/admin/solr-admin.css +206 -0
- data/lib/webapp/admin/solr-head.gif +0 -0
- data/lib/webapp/admin/solr-head.png +0 -0
- data/lib/webapp/admin/solr-lowercase.gif +0 -0
- data/lib/webapp/admin/solr-lowercase.png +0 -0
- data/lib/webapp/admin/stats.jsp +94 -0
- data/lib/webapp/admin/stats.xsl +220 -0
- data/lib/webapp/admin/tabular.xsl +141 -0
- data/lib/webapp/admin/threaddump.jsp +110 -0
- data/lib/webapp/admin/threaddump.xsl +103 -0
- data/lib/webapp/favicon.ico +0 -0
- data/lib/webapp/index.jsp +42 -0
- metadata +120 -0
@@ -0,0 +1,110 @@
|
|
1
|
+
<%--
|
2
|
+
Licensed to the Apache Software Foundation (ASF) under one or more
|
3
|
+
contributor license agreements. See the NOTICE file distributed with
|
4
|
+
this work for additional information regarding copyright ownership.
|
5
|
+
The ASF licenses this file to You under the Apache License, Version 2.0
|
6
|
+
(the "License"); you may not use this file except in compliance with
|
7
|
+
the License. You may obtain a copy of the License at
|
8
|
+
|
9
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
10
|
+
|
11
|
+
Unless required by applicable law or agreed to in writing, software
|
12
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
13
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
14
|
+
See the License for the specific language governing permissions and
|
15
|
+
limitations under the License.
|
16
|
+
--%>
|
17
|
+
<%@ page import="org.apache.solr.core.SolrConfig,
|
18
|
+
org.apache.solr.core.SolrCore,
|
19
|
+
org.apache.solr.schema.IndexSchema,
|
20
|
+
java.io.File"%>
|
21
|
+
<%@ page import="java.net.InetAddress"%>
|
22
|
+
<%@ page import="java.io.StringWriter"%>
|
23
|
+
<%@ page import="org.apache.solr.core.Config"%>
|
24
|
+
<%@ page import="org.apache.solr.common.util.XML"%>
|
25
|
+
<%@ page import="org.apache.solr.common.SolrException"%>
|
26
|
+
<%@ page import="org.apache.lucene.LucenePackage"%>
|
27
|
+
<%@ page import="java.net.UnknownHostException" %>
|
28
|
+
<%
|
29
|
+
//
|
30
|
+
SolrCore core = (SolrCore) request.getAttribute("org.apache.solr.SolrCore");
|
31
|
+
if (core == null) {
|
32
|
+
response.sendError( 404, "missing core name in path" );
|
33
|
+
return;
|
34
|
+
}
|
35
|
+
|
36
|
+
SolrConfig solrConfig = core.getSolrConfig();
|
37
|
+
int port = request.getServerPort();
|
38
|
+
IndexSchema schema = core.getSchema();
|
39
|
+
|
40
|
+
// enabled/disabled is purely from the point of a load-balancer
|
41
|
+
// and has no effect on local server function. If there is no healthcheck
|
42
|
+
// configured, don't put any status on the admin pages.
|
43
|
+
String enabledStatus = null;
|
44
|
+
String enabledFile = solrConfig.get("admin/healthcheck/text()",null);
|
45
|
+
boolean isEnabled = false;
|
46
|
+
if (enabledFile!=null) {
|
47
|
+
isEnabled = new File(enabledFile).exists();
|
48
|
+
}
|
49
|
+
|
50
|
+
String collectionName = schema!=null ? schema.getName():"unknown";
|
51
|
+
InetAddress addr = null;
|
52
|
+
String hostname = "unknown";
|
53
|
+
try {
|
54
|
+
addr = InetAddress.getLocalHost();
|
55
|
+
hostname = addr.getCanonicalHostName();
|
56
|
+
} catch (UnknownHostException e) {
|
57
|
+
//default to unknown
|
58
|
+
}
|
59
|
+
|
60
|
+
String defaultSearch = "";
|
61
|
+
{
|
62
|
+
StringWriter tmp = new StringWriter();
|
63
|
+
XML.escapeCharData
|
64
|
+
(solrConfig.get("admin/defaultQuery/text()", null), tmp);
|
65
|
+
defaultSearch = tmp.toString();
|
66
|
+
}
|
67
|
+
|
68
|
+
String solrImplVersion = "";
|
69
|
+
String solrSpecVersion = "";
|
70
|
+
String luceneImplVersion = "";
|
71
|
+
String luceneSpecVersion = "";
|
72
|
+
|
73
|
+
{
|
74
|
+
Package p;
|
75
|
+
StringWriter tmp;
|
76
|
+
|
77
|
+
p = SolrCore.class.getPackage();
|
78
|
+
|
79
|
+
tmp = new StringWriter();
|
80
|
+
solrImplVersion = p.getImplementationVersion();
|
81
|
+
if (null != solrImplVersion) {
|
82
|
+
XML.escapeCharData(solrImplVersion, tmp);
|
83
|
+
solrImplVersion = tmp.toString();
|
84
|
+
}
|
85
|
+
tmp = new StringWriter();
|
86
|
+
solrSpecVersion = p.getSpecificationVersion() ;
|
87
|
+
if (null != solrSpecVersion) {
|
88
|
+
XML.escapeCharData(solrSpecVersion, tmp);
|
89
|
+
solrSpecVersion = tmp.toString();
|
90
|
+
}
|
91
|
+
|
92
|
+
p = LucenePackage.class.getPackage();
|
93
|
+
|
94
|
+
tmp = new StringWriter();
|
95
|
+
luceneImplVersion = p.getImplementationVersion();
|
96
|
+
if (null != luceneImplVersion) {
|
97
|
+
XML.escapeCharData(luceneImplVersion, tmp);
|
98
|
+
luceneImplVersion = tmp.toString();
|
99
|
+
}
|
100
|
+
tmp = new StringWriter();
|
101
|
+
luceneSpecVersion = p.getSpecificationVersion() ;
|
102
|
+
if (null != luceneSpecVersion) {
|
103
|
+
XML.escapeCharData(luceneSpecVersion, tmp);
|
104
|
+
luceneSpecVersion = tmp.toString();
|
105
|
+
}
|
106
|
+
}
|
107
|
+
|
108
|
+
String cwd=System.getProperty("user.dir");
|
109
|
+
String solrHome= solrConfig.getInstanceDir();
|
110
|
+
%>
|
@@ -0,0 +1,116 @@
|
|
1
|
+
<%@ page contentType="text/html; charset=utf-8" pageEncoding="UTF-8"%>
|
2
|
+
<%--
|
3
|
+
Licensed to the Apache Software Foundation (ASF) under one or more
|
4
|
+
contributor license agreements. See the NOTICE file distributed with
|
5
|
+
this work for additional information regarding copyright ownership.
|
6
|
+
The ASF licenses this file to You under the Apache License, Version 2.0
|
7
|
+
(the "License"); you may not use this file except in compliance with
|
8
|
+
the License. You may obtain a copy of the License at
|
9
|
+
|
10
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
|
12
|
+
Unless required by applicable law or agreed to in writing, software
|
13
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
14
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
15
|
+
See the License for the specific language governing permissions and
|
16
|
+
limitations under the License.
|
17
|
+
--%>
|
18
|
+
<%@ page import="org.apache.solr.common.util.XML"%>
|
19
|
+
<%@ page import="org.apache.solr.core.SolrCore,
|
20
|
+
org.apache.solr.schema.IndexSchema,
|
21
|
+
java.io.File,
|
22
|
+
java.net.InetAddress,
|
23
|
+
java.net.UnknownHostException"%>
|
24
|
+
<%@ page import="java.util.Date"%>
|
25
|
+
<%@ page import="java.util.logging.Level"%>
|
26
|
+
<%@ page import="java.util.logging.Logger"%>
|
27
|
+
<%@include file="header.jsp" %>
|
28
|
+
<%
|
29
|
+
|
30
|
+
String action = request.getParameter("action");
|
31
|
+
String logging = request.getParameter("log");
|
32
|
+
String enableActionStatus = "";
|
33
|
+
boolean isValid = false;
|
34
|
+
|
35
|
+
if (action != null) {
|
36
|
+
// Validate fname
|
37
|
+
if ("Enable".compareTo(action) == 0) isValid = true;
|
38
|
+
if ("Disable".compareTo(action) == 0) isValid = true;
|
39
|
+
}
|
40
|
+
if (logging != null) {
|
41
|
+
action = "Set Log Level";
|
42
|
+
isValid = true;
|
43
|
+
}
|
44
|
+
if (isValid) {
|
45
|
+
if ("Enable".compareTo(action) == 0) {
|
46
|
+
try {
|
47
|
+
File enableFile = new File(enabledFile);
|
48
|
+
if (enableFile.createNewFile()) {
|
49
|
+
enableActionStatus += "Enable Succeeded (enable file ";
|
50
|
+
enableActionStatus += enabledFile;
|
51
|
+
enableActionStatus += " created)";
|
52
|
+
} else {
|
53
|
+
enableActionStatus += "Already Enabled";
|
54
|
+
}
|
55
|
+
} catch(Exception e) {
|
56
|
+
enableActionStatus += "Enable Failed: " + e.toString();
|
57
|
+
}
|
58
|
+
}
|
59
|
+
if ("Disable".compareTo(action) == 0) {
|
60
|
+
try {
|
61
|
+
File enableFile = new File(enabledFile);
|
62
|
+
if (enableFile.delete()) {
|
63
|
+
enableActionStatus = "Disable Succeeded (enable file ";
|
64
|
+
enableActionStatus += enabledFile;
|
65
|
+
enableActionStatus += " removed)";
|
66
|
+
} else {
|
67
|
+
enableActionStatus = "Already Disabled";
|
68
|
+
}
|
69
|
+
} catch(Exception e) {
|
70
|
+
enableActionStatus += "Disable Failed: " + e.toString();
|
71
|
+
}
|
72
|
+
}
|
73
|
+
if (logging != null) {
|
74
|
+
try {
|
75
|
+
Logger log = SolrCore.log;
|
76
|
+
Logger parent = log.getParent();
|
77
|
+
while (parent != null) {
|
78
|
+
log = parent;
|
79
|
+
parent = log.getParent();
|
80
|
+
}
|
81
|
+
log.setLevel(Level.parse(logging));
|
82
|
+
enableActionStatus = "Set Log Level (" + logging + ") Succeeded";
|
83
|
+
} catch(Exception e) {
|
84
|
+
enableActionStatus += "Set Log Level (" + logging + ") Failed: "
|
85
|
+
+ e.toString();
|
86
|
+
}
|
87
|
+
}
|
88
|
+
} else {
|
89
|
+
enableActionStatus = "Illegal Action";
|
90
|
+
}
|
91
|
+
// :TODO: might want to rework this so any logging change happens *after*
|
92
|
+
SolrCore.log.log(Level.INFO, enableActionStatus);
|
93
|
+
%>
|
94
|
+
<br clear="all">
|
95
|
+
<table>
|
96
|
+
<tr>
|
97
|
+
<td>
|
98
|
+
<H3>Action:</H3>
|
99
|
+
</td>
|
100
|
+
<td>
|
101
|
+
<% XML.escapeCharData(action, out); %><br>
|
102
|
+
</td>
|
103
|
+
</tr>
|
104
|
+
<tr>
|
105
|
+
<td>
|
106
|
+
<H4>Result:</H4>
|
107
|
+
</td>
|
108
|
+
<td>
|
109
|
+
<% XML.escapeCharData(enableActionStatus, out); %><br>
|
110
|
+
</td>
|
111
|
+
</tr>
|
112
|
+
</table>
|
113
|
+
<br><br>
|
114
|
+
<a href=".">Return to Admin Page</a>
|
115
|
+
</body>
|
116
|
+
</html>
|
@@ -0,0 +1,456 @@
|
|
1
|
+
<%@ page contentType="text/html; charset=utf-8" pageEncoding="UTF-8"%>
|
2
|
+
<%--
|
3
|
+
Licensed to the Apache Software Foundation (ASF) under one or more
|
4
|
+
contributor license agreements. See the NOTICE file distributed with
|
5
|
+
this work for additional information regarding copyright ownership.
|
6
|
+
The ASF licenses this file to You under the Apache License, Version 2.0
|
7
|
+
(the "License"); you may not use this file except in compliance with
|
8
|
+
the License. You may obtain a copy of the License at
|
9
|
+
|
10
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
|
12
|
+
Unless required by applicable law or agreed to in writing, software
|
13
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
14
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
15
|
+
See the License for the specific language governing permissions and
|
16
|
+
limitations under the License.
|
17
|
+
--%>
|
18
|
+
<%@ page import="org.apache.lucene.analysis.Analyzer,
|
19
|
+
org.apache.lucene.analysis.Token,
|
20
|
+
org.apache.lucene.analysis.TokenStream,
|
21
|
+
org.apache.lucene.index.Payload,
|
22
|
+
org.apache.solr.analysis.TokenFilterFactory,
|
23
|
+
org.apache.solr.analysis.TokenizerChain,
|
24
|
+
org.apache.solr.analysis.TokenizerFactory,
|
25
|
+
org.apache.solr.schema.FieldType,
|
26
|
+
org.apache.solr.schema.SchemaField,
|
27
|
+
org.apache.solr.common.util.XML,
|
28
|
+
javax.servlet.jsp.JspWriter,java.io.IOException
|
29
|
+
"%>
|
30
|
+
<%@ page import="java.io.Reader"%>
|
31
|
+
<%@ page import="java.io.StringReader"%>
|
32
|
+
<%@ page import="java.util.*"%>
|
33
|
+
<%@ page import="java.math.BigInteger" %>
|
34
|
+
|
35
|
+
<%-- $Id: analysis.jsp 677047 2008-07-15 21:30:37Z hossman $ --%>
|
36
|
+
<%-- $Source: /cvs/main/searching/org.apache.solrolarServer/resources/admin/analysis.jsp,v $ --%>
|
37
|
+
<%-- $Name: $ --%>
|
38
|
+
|
39
|
+
<%@include file="header.jsp" %>
|
40
|
+
|
41
|
+
<%
|
42
|
+
// is name a field name or a type name?
|
43
|
+
String nt = request.getParameter("nt");
|
44
|
+
if (nt==null || nt.length()==0) nt="name"; // assume field name
|
45
|
+
nt = nt.toLowerCase().trim();
|
46
|
+
String name = request.getParameter("name");
|
47
|
+
if (name==null || name.length()==0) name="";
|
48
|
+
String val = request.getParameter("val");
|
49
|
+
if (val==null || val.length()==0) val="";
|
50
|
+
String qval = request.getParameter("qval");
|
51
|
+
if (qval==null || qval.length()==0) qval="";
|
52
|
+
String verboseS = request.getParameter("verbose");
|
53
|
+
boolean verbose = verboseS!=null && verboseS.equalsIgnoreCase("on");
|
54
|
+
String qverboseS = request.getParameter("qverbose");
|
55
|
+
boolean qverbose = qverboseS!=null && qverboseS.equalsIgnoreCase("on");
|
56
|
+
String highlightS = request.getParameter("highlight");
|
57
|
+
boolean highlight = highlightS!=null && highlightS.equalsIgnoreCase("on");
|
58
|
+
%>
|
59
|
+
|
60
|
+
<br clear="all">
|
61
|
+
|
62
|
+
<h2>Field Analysis</h2>
|
63
|
+
|
64
|
+
<form method="POST" action="analysis.jsp" accept-charset="UTF-8">
|
65
|
+
<table>
|
66
|
+
<tr>
|
67
|
+
<td>
|
68
|
+
<strong>Field
|
69
|
+
<select name="nt">
|
70
|
+
<option <%= nt.equals("name") ? "selected=\"selected\"" : "" %> >name</option>
|
71
|
+
<option <%= nt.equals("type") ? "selected=\"selected\"" : "" %>>type</option>
|
72
|
+
</select></strong>
|
73
|
+
</td>
|
74
|
+
<td>
|
75
|
+
<input class="std" name="name" type="text" value="<% XML.escapeCharData(name, out); %>">
|
76
|
+
</td>
|
77
|
+
</tr>
|
78
|
+
<tr>
|
79
|
+
<td>
|
80
|
+
<strong>Field value (Index)</strong>
|
81
|
+
<br/>
|
82
|
+
verbose output
|
83
|
+
<input name="verbose" type="checkbox"
|
84
|
+
<%= verbose ? "checked=\"true\"" : "" %> >
|
85
|
+
<br/>
|
86
|
+
highlight matches
|
87
|
+
<input name="highlight" type="checkbox"
|
88
|
+
<%= highlight ? "checked=\"true\"" : "" %> >
|
89
|
+
</td>
|
90
|
+
<td>
|
91
|
+
<textarea class="std" rows="8" cols="70" name="val"><% XML.escapeCharData(val,out); %></textarea>
|
92
|
+
</td>
|
93
|
+
</tr>
|
94
|
+
<tr>
|
95
|
+
<td>
|
96
|
+
<strong>Field value (Query)</strong>
|
97
|
+
<br/>
|
98
|
+
verbose output
|
99
|
+
<input name="qverbose" type="checkbox"
|
100
|
+
<%= qverbose ? "checked=\"true\"" : "" %> >
|
101
|
+
</td>
|
102
|
+
<td>
|
103
|
+
<textarea class="std" rows="1" cols="70" name="qval"><% XML.escapeCharData(qval,out); %></textarea>
|
104
|
+
</td>
|
105
|
+
</tr>
|
106
|
+
<tr>
|
107
|
+
|
108
|
+
<td>
|
109
|
+
</td>
|
110
|
+
|
111
|
+
<td>
|
112
|
+
<input class="stdbutton" type="submit" value="analyze">
|
113
|
+
</td>
|
114
|
+
|
115
|
+
</tr>
|
116
|
+
</table>
|
117
|
+
</form>
|
118
|
+
|
119
|
+
|
120
|
+
<%
|
121
|
+
SchemaField field=null;
|
122
|
+
|
123
|
+
if (name!="") {
|
124
|
+
if (nt.equals("name")) {
|
125
|
+
try {
|
126
|
+
field = schema.getField(name);
|
127
|
+
} catch (Exception e) {
|
128
|
+
out.print("<strong>Unknown Field: ");
|
129
|
+
XML.escapeCharData(name, out);
|
130
|
+
out.println("</strong>");
|
131
|
+
}
|
132
|
+
} else {
|
133
|
+
FieldType t = schema.getFieldTypes().get(name);
|
134
|
+
if (null == t) {
|
135
|
+
out.print("<strong>Unknown Field Type: ");
|
136
|
+
XML.escapeCharData(name, out);
|
137
|
+
out.println("</strong>");
|
138
|
+
} else {
|
139
|
+
field = new SchemaField("fakefieldoftype:"+name, t);
|
140
|
+
}
|
141
|
+
}
|
142
|
+
}
|
143
|
+
|
144
|
+
if (field!=null) {
|
145
|
+
HashSet<Tok> matches = null;
|
146
|
+
if (qval!="" && highlight) {
|
147
|
+
Reader reader = new StringReader(qval);
|
148
|
+
Analyzer analyzer = field.getType().getQueryAnalyzer();
|
149
|
+
TokenStream tstream = analyzer.tokenStream(field.getName(),reader);
|
150
|
+
List<Token> tokens = getTokens(tstream);
|
151
|
+
matches = new HashSet<Tok>();
|
152
|
+
for (Token t : tokens) { matches.add( new Tok(t,0)); }
|
153
|
+
}
|
154
|
+
|
155
|
+
if (val!="") {
|
156
|
+
out.println("<h3>Index Analyzer</h3>");
|
157
|
+
doAnalyzer(out, field, val, false, verbose,matches);
|
158
|
+
}
|
159
|
+
if (qval!="") {
|
160
|
+
out.println("<h3>Query Analyzer</h3>");
|
161
|
+
doAnalyzer(out, field, qval, true, qverbose,null);
|
162
|
+
}
|
163
|
+
}
|
164
|
+
|
165
|
+
%>
|
166
|
+
|
167
|
+
|
168
|
+
</body>
|
169
|
+
</html>
|
170
|
+
|
171
|
+
|
172
|
+
<%!
|
173
|
+
private static void doAnalyzer(JspWriter out, SchemaField field, String val, boolean queryAnalyser, boolean verbose, Set<Tok> match) throws Exception {
|
174
|
+
Reader reader = new StringReader(val);
|
175
|
+
|
176
|
+
FieldType ft = field.getType();
|
177
|
+
Analyzer analyzer = queryAnalyser ?
|
178
|
+
ft.getQueryAnalyzer() : ft.getAnalyzer();
|
179
|
+
if (analyzer instanceof TokenizerChain) {
|
180
|
+
TokenizerChain tchain = (TokenizerChain)analyzer;
|
181
|
+
TokenizerFactory tfac = tchain.getTokenizerFactory();
|
182
|
+
TokenFilterFactory[] filtfacs = tchain.getTokenFilterFactories();
|
183
|
+
|
184
|
+
TokenStream tstream = tfac.create(reader);
|
185
|
+
List<Token> tokens = getTokens(tstream);
|
186
|
+
tstream = tfac.create(reader);
|
187
|
+
if (verbose) {
|
188
|
+
writeHeader(out, tfac.getClass(), tfac.getArgs());
|
189
|
+
}
|
190
|
+
|
191
|
+
writeTokens(out, tokens, ft, verbose, match);
|
192
|
+
|
193
|
+
for (TokenFilterFactory filtfac : filtfacs) {
|
194
|
+
if (verbose) {
|
195
|
+
writeHeader(out, filtfac.getClass(), filtfac.getArgs());
|
196
|
+
}
|
197
|
+
|
198
|
+
final Iterator<Token> iter = tokens.iterator();
|
199
|
+
tstream = filtfac.create( new TokenStream() {
|
200
|
+
public Token next() throws IOException {
|
201
|
+
return iter.hasNext() ? iter.next() : null;
|
202
|
+
}
|
203
|
+
}
|
204
|
+
);
|
205
|
+
tokens = getTokens(tstream);
|
206
|
+
|
207
|
+
writeTokens(out, tokens, ft, verbose, match);
|
208
|
+
}
|
209
|
+
|
210
|
+
} else {
|
211
|
+
TokenStream tstream = analyzer.tokenStream(field.getName(),reader);
|
212
|
+
List<Token> tokens = getTokens(tstream);
|
213
|
+
if (verbose) {
|
214
|
+
writeHeader(out, analyzer.getClass(), new HashMap<String,String>());
|
215
|
+
}
|
216
|
+
writeTokens(out, tokens, ft, verbose, match);
|
217
|
+
}
|
218
|
+
}
|
219
|
+
|
220
|
+
|
221
|
+
static List<Token> getTokens(TokenStream tstream) throws IOException {
|
222
|
+
List<Token> tokens = new ArrayList<Token>();
|
223
|
+
while (true) {
|
224
|
+
Token t = tstream.next();
|
225
|
+
if (t==null) break;
|
226
|
+
tokens.add(t);
|
227
|
+
}
|
228
|
+
return tokens;
|
229
|
+
}
|
230
|
+
|
231
|
+
|
232
|
+
private static class Tok {
|
233
|
+
Token token;
|
234
|
+
int pos;
|
235
|
+
Tok(Token token, int pos) {
|
236
|
+
this.token=token;
|
237
|
+
this.pos=pos;
|
238
|
+
}
|
239
|
+
|
240
|
+
public boolean equals(Object o) {
|
241
|
+
return ((Tok)o).token.termText().equals(token.termText());
|
242
|
+
}
|
243
|
+
public int hashCode() {
|
244
|
+
return token.termText().hashCode();
|
245
|
+
}
|
246
|
+
public String toString() {
|
247
|
+
return token.termText();
|
248
|
+
}
|
249
|
+
}
|
250
|
+
|
251
|
+
private static interface ToStr {
|
252
|
+
public String toStr(Object o);
|
253
|
+
}
|
254
|
+
|
255
|
+
private static void printRow(JspWriter out, String header, List[] arrLst, ToStr converter, boolean multival, boolean verbose, Set<Tok> match) throws IOException {
|
256
|
+
// find the maximum number of terms for any position
|
257
|
+
int maxSz=1;
|
258
|
+
if (multival) {
|
259
|
+
for (List lst : arrLst) {
|
260
|
+
maxSz = Math.max(lst.size(), maxSz);
|
261
|
+
}
|
262
|
+
}
|
263
|
+
|
264
|
+
|
265
|
+
for (int idx=0; idx<maxSz; idx++) {
|
266
|
+
out.println("<tr>");
|
267
|
+
if (idx==0 && verbose) {
|
268
|
+
if (header != null) {
|
269
|
+
out.print("<th NOWRAP rowspan=\""+maxSz+"\">");
|
270
|
+
XML.escapeCharData(header,out);
|
271
|
+
out.println("</th>");
|
272
|
+
}
|
273
|
+
}
|
274
|
+
|
275
|
+
for (int posIndex=0; posIndex<arrLst.length; posIndex++) {
|
276
|
+
List<Tok> lst = arrLst[posIndex];
|
277
|
+
if (lst.size() <= idx) continue;
|
278
|
+
if (match!=null && match.contains(lst.get(idx))) {
|
279
|
+
out.print("<td class=\"highlight\"");
|
280
|
+
} else {
|
281
|
+
out.print("<td class=\"debugdata\"");
|
282
|
+
}
|
283
|
+
|
284
|
+
// if the last value in the column, use up
|
285
|
+
// the rest of the space via rowspan.
|
286
|
+
if (lst.size() == idx+1 && lst.size() < maxSz) {
|
287
|
+
out.print("rowspan=\""+(maxSz-lst.size()+1)+'"');
|
288
|
+
}
|
289
|
+
|
290
|
+
out.print('>');
|
291
|
+
|
292
|
+
XML.escapeCharData(converter.toStr(lst.get(idx)), out);
|
293
|
+
out.print("</td>");
|
294
|
+
}
|
295
|
+
|
296
|
+
out.println("</tr>");
|
297
|
+
}
|
298
|
+
|
299
|
+
}
|
300
|
+
|
301
|
+
static String isPayloadString( Payload p ) {
|
302
|
+
String sp = new String( p.getData() );
|
303
|
+
for( int i=0; i < sp.length(); i++ ) {
|
304
|
+
if( !Character.isDefined( sp.charAt(i) ) || Character.isISOControl( sp.charAt(i) ) )
|
305
|
+
return "";
|
306
|
+
}
|
307
|
+
return "(" + sp + ")";
|
308
|
+
}
|
309
|
+
|
310
|
+
static void writeHeader(JspWriter out, Class clazz, Map<String,String> args) throws IOException {
|
311
|
+
out.print("<h4>");
|
312
|
+
out.print(clazz.getName());
|
313
|
+
XML.escapeCharData(" "+args,out);
|
314
|
+
out.println("</h4>");
|
315
|
+
}
|
316
|
+
|
317
|
+
|
318
|
+
|
319
|
+
// readable, raw, pos, type, start/end
|
320
|
+
static void writeTokens(JspWriter out, List<Token> tokens, final FieldType ft, boolean verbose, Set<Tok> match) throws IOException {
|
321
|
+
|
322
|
+
// Use a map to tell what tokens are in what positions
|
323
|
+
// because some tokenizers/filters may do funky stuff with
|
324
|
+
// very large increments, or negative increments.
|
325
|
+
HashMap<Integer,List<Tok>> map = new HashMap<Integer,List<Tok>>();
|
326
|
+
boolean needRaw=false;
|
327
|
+
int pos=0;
|
328
|
+
for (Token t : tokens) {
|
329
|
+
if (!t.termText().equals(ft.indexedToReadable(t.termText()))) {
|
330
|
+
needRaw=true;
|
331
|
+
}
|
332
|
+
|
333
|
+
pos += t.getPositionIncrement();
|
334
|
+
List lst = map.get(pos);
|
335
|
+
if (lst==null) {
|
336
|
+
lst = new ArrayList(1);
|
337
|
+
map.put(pos,lst);
|
338
|
+
}
|
339
|
+
Tok tok = new Tok(t,pos);
|
340
|
+
lst.add(tok);
|
341
|
+
}
|
342
|
+
|
343
|
+
List<Tok>[] arr = (List<Tok>[])map.values().toArray(new ArrayList[map.size()]);
|
344
|
+
|
345
|
+
/* Jetty 6.1.3 miscompiles this generics version...
|
346
|
+
Arrays.sort(arr, new Comparator<List<Tok>>() {
|
347
|
+
public int compare(List<Tok> toks, List<Tok> toks1) {
|
348
|
+
return toks.get(0).pos - toks1.get(0).pos;
|
349
|
+
}
|
350
|
+
}
|
351
|
+
*/
|
352
|
+
|
353
|
+
Arrays.sort(arr, new Comparator() {
|
354
|
+
public int compare(Object toks, Object toks1) {
|
355
|
+
return ((List<Tok>)toks).get(0).pos - ((List<Tok>)toks1).get(0).pos;
|
356
|
+
}
|
357
|
+
}
|
358
|
+
|
359
|
+
|
360
|
+
);
|
361
|
+
|
362
|
+
out.println("<table width=\"auto\" class=\"analysis\" border=\"1\">");
|
363
|
+
|
364
|
+
if (verbose) {
|
365
|
+
printRow(out,"term position", arr, new ToStr() {
|
366
|
+
public String toStr(Object o) {
|
367
|
+
return Integer.toString(((Tok)o).pos);
|
368
|
+
}
|
369
|
+
}
|
370
|
+
,false
|
371
|
+
,verbose
|
372
|
+
,null);
|
373
|
+
}
|
374
|
+
|
375
|
+
|
376
|
+
printRow(out,"term text", arr, new ToStr() {
|
377
|
+
public String toStr(Object o) {
|
378
|
+
return ft.indexedToReadable( ((Tok)o).token.termText() );
|
379
|
+
}
|
380
|
+
}
|
381
|
+
,true
|
382
|
+
,verbose
|
383
|
+
,match
|
384
|
+
);
|
385
|
+
|
386
|
+
if (needRaw) {
|
387
|
+
printRow(out,"raw text", arr, new ToStr() {
|
388
|
+
public String toStr(Object o) {
|
389
|
+
// page is UTF-8, so anything goes.
|
390
|
+
return ((Tok)o).token.termText();
|
391
|
+
}
|
392
|
+
}
|
393
|
+
,true
|
394
|
+
,verbose
|
395
|
+
,match
|
396
|
+
);
|
397
|
+
}
|
398
|
+
|
399
|
+
if (verbose) {
|
400
|
+
printRow(out,"term type", arr, new ToStr() {
|
401
|
+
public String toStr(Object o) {
|
402
|
+
String tt = ((Tok)o).token.type();
|
403
|
+
if (tt == null) {
|
404
|
+
return "null";
|
405
|
+
} else {
|
406
|
+
return tt;
|
407
|
+
}
|
408
|
+
}
|
409
|
+
}
|
410
|
+
,true
|
411
|
+
,verbose,
|
412
|
+
null
|
413
|
+
);
|
414
|
+
}
|
415
|
+
|
416
|
+
if (verbose) {
|
417
|
+
printRow(out,"source start,end", arr, new ToStr() {
|
418
|
+
public String toStr(Object o) {
|
419
|
+
Token t = ((Tok)o).token;
|
420
|
+
return Integer.toString(t.startOffset()) + ',' + t.endOffset() ;
|
421
|
+
}
|
422
|
+
}
|
423
|
+
,true
|
424
|
+
,verbose
|
425
|
+
,null
|
426
|
+
);
|
427
|
+
}
|
428
|
+
|
429
|
+
if (verbose) {
|
430
|
+
printRow(out,"payload", arr, new ToStr() {
|
431
|
+
public String toStr(Object o) {
|
432
|
+
Token t = ((Tok)o).token;
|
433
|
+
Payload p = t.getPayload();
|
434
|
+
if( null != p ) {
|
435
|
+
BigInteger bi = new BigInteger( p.getData() );
|
436
|
+
String ret = bi.toString( 16 );
|
437
|
+
if (ret.length() % 2 != 0) {
|
438
|
+
// Pad with 0
|
439
|
+
ret = "0"+ret;
|
440
|
+
}
|
441
|
+
ret += isPayloadString( p );
|
442
|
+
return ret;
|
443
|
+
}
|
444
|
+
return "";
|
445
|
+
}
|
446
|
+
}
|
447
|
+
,true
|
448
|
+
,verbose
|
449
|
+
,null
|
450
|
+
);
|
451
|
+
}
|
452
|
+
|
453
|
+
out.println("</table>");
|
454
|
+
}
|
455
|
+
|
456
|
+
%>
|